TY - JOUR
T1 - Real-time processing methods to characterize streamwise vortices
AU - Braud, Caroline
AU - Liberzon, Alex
N1 - Publisher Copyright:
© 2018 Elsevier Ltd
PY - 2018/8
Y1 - 2018/8
N2 - One of the bottlenecks of an active control of turbulent boundary layers based on particle image velocimetry (PIV) measurements relates to the ability to process PIV images and extract useful for control information in real-time, on a time scale relevant for the selected control strategy. We propose two methodologies to extract useful information from PIV measurements in the y−z spanwise - wall normal cross-section, based on characterisation of properties such as size and strength of streamwise vortices. The vortices are created by active vortex generators embedded in a fully turbulent boundary layer. The proposed methodologies combine vortex identification and characterisation methods that use instantaneous PIV realisations to extract centers and strengths of streamwise vortices. The main purpose is to compare a standard vortex identification method that provides a robust and accurate estimate with another, ad-hoc method that is less robust or accurate but has a large computational speed gain potential. For demonstration purposes we use PIV measurements obtained at the wind tunnel facility downstream active vortex generators (Carlier and Stanislas, 2005; Foucaut et al., 2014). The first algorithm uses the Q-criteria and the integration of vorticity of each extracted vortex. This robust and accurate method requires the full spatially resolved PIV field followed by the computationally expensive spatial derivatives calculations and an integration. The second algorithm utilises the prior knowledge about the presence and shape of streamwise vortices expected in the measurement, therefore uses only a single line velocity profile estimate and can be parallelized to obtain few horizontal velocity profiles at different wall-normal distances. We compare the two methods and discuss their potential in term of computation times and robustness. Results show that in a turbulent boundary layer, a moving window average of a number of instantaneous fields is needed for both methods.
AB - One of the bottlenecks of an active control of turbulent boundary layers based on particle image velocimetry (PIV) measurements relates to the ability to process PIV images and extract useful for control information in real-time, on a time scale relevant for the selected control strategy. We propose two methodologies to extract useful information from PIV measurements in the y−z spanwise - wall normal cross-section, based on characterisation of properties such as size and strength of streamwise vortices. The vortices are created by active vortex generators embedded in a fully turbulent boundary layer. The proposed methodologies combine vortex identification and characterisation methods that use instantaneous PIV realisations to extract centers and strengths of streamwise vortices. The main purpose is to compare a standard vortex identification method that provides a robust and accurate estimate with another, ad-hoc method that is less robust or accurate but has a large computational speed gain potential. For demonstration purposes we use PIV measurements obtained at the wind tunnel facility downstream active vortex generators (Carlier and Stanislas, 2005; Foucaut et al., 2014). The first algorithm uses the Q-criteria and the integration of vorticity of each extracted vortex. This robust and accurate method requires the full spatially resolved PIV field followed by the computationally expensive spatial derivatives calculations and an integration. The second algorithm utilises the prior knowledge about the presence and shape of streamwise vortices expected in the measurement, therefore uses only a single line velocity profile estimate and can be parallelized to obtain few horizontal velocity profiles at different wall-normal distances. We compare the two methods and discuss their potential in term of computation times and robustness. Results show that in a turbulent boundary layer, a moving window average of a number of instantaneous fields is needed for both methods.
KW - Active vortex generator
KW - Boundary layer control
KW - Particle image velocimetry
KW - Vortex identification
UR - http://www.scopus.com/inward/record.url?scp=85047256729&partnerID=8YFLogxK
U2 - 10.1016/j.jweia.2018.05.006
DO - 10.1016/j.jweia.2018.05.006
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
AN - SCOPUS:85047256729
SN - 0167-6105
VL - 179
SP - 14
EP - 25
JO - Journal of Wind Engineering and Industrial Aerodynamics
JF - Journal of Wind Engineering and Industrial Aerodynamics
ER -